Mailing-List: contact dev-help@tephra.incubator.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@tephra.incubator.apache.org
Date: Thu, 9 Feb 2017 00:07:42 +0000 (UTC)
From: "ASF GitHub Bot (JIRA)" <jira@apache.org>
To: dev@tephra.incubator.apache.org
Message-ID: <JIRA.13039995.1486079998000.33645.1486598862089@Atlassian.JIRA>
In-Reply-To: <JIRA.13039995.1486079998000@Atlassian.JIRA>
References: <JIRA.13039995.1486079998000@Atlassian.JIRA> <JIRA.13039995.1486079998700@jira-lw-us.apache.org>
Subject: [jira] [Commented] (TEPHRA-215) Share PruneUpperBoundWriter across
 all TransactionProcessors on the same region server
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit
archived-at: Thu, 09 Feb 2017 00:07:54 -0000


    [ https://issues.apache.org/jira/browse/TEPHRA-215?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15858737#comment-15858737 ] 

ASF GitHub Bot commented on TEPHRA-215:
---------------------------------------

Github user poornachandra commented on a diff in the pull request:

    https://github.com/apache/incubator-tephra/pull/32#discussion_r100201530
  
    --- Diff: tephra-hbase-compat-1.1-base/src/main/java/org/apache/tephra/hbase/txprune/PruneUpperBoundWriter.java ---
    @@ -18,77 +18,85 @@
     
     package org.apache.tephra.hbase.txprune;
     
    +import com.google.common.util.concurrent.AbstractIdleService;
     import org.apache.commons.logging.Log;
     import org.apache.commons.logging.LogFactory;
     import org.apache.hadoop.hbase.TableName;
    -import org.apache.tephra.hbase.coprocessor.TransactionProcessor;
     
     import java.io.IOException;
    +import java.nio.ByteBuffer;
    +import java.util.Map;
    +import java.util.concurrent.ConcurrentSkipListMap;
     import java.util.concurrent.TimeUnit;
    -import java.util.concurrent.atomic.AtomicBoolean;
    -import java.util.concurrent.atomic.AtomicLong;
     
     /**
      * Thread that will write the the prune upper bound
      */
    -public class PruneUpperBoundWriter {
    -  private static final Log LOG = LogFactory.getLog(TransactionProcessor.class);
    +public class PruneUpperBoundWriter extends AbstractIdleService {
    +  private static final Log LOG = LogFactory.getLog(PruneUpperBoundWriter.class);
     
    -  private final TableName pruneStateTable;
    +  private final TableName tableName;
       private final DataJanitorState dataJanitorState;
    -  private final byte[] regionName;
    -  private final String regionNameAsString;
       private final long pruneFlushInterval;
    -  private final AtomicLong pruneUpperBound;
    -  private final AtomicBoolean shouldFlush;
    +  private final ConcurrentSkipListMap<ByteBuffer, Long> pruneEntries;
    +
    +  private volatile boolean stopped;
    +  private volatile Thread flushThread;
     
    -  private Thread flushThread;
       private long lastChecked;
     
    -  public PruneUpperBoundWriter(DataJanitorState dataJanitorState, TableName pruneStateTable, String regionNameAsString,
    -                               byte[] regionName, long pruneFlushInterval) {
    -    this.pruneStateTable = pruneStateTable;
    +  public PruneUpperBoundWriter(TableName tableName, DataJanitorState dataJanitorState, long pruneFlushInterval) {
    +    this.tableName = tableName;
         this.dataJanitorState = dataJanitorState;
    -    this.regionName = regionName;
    -    this.regionNameAsString = regionNameAsString;
         this.pruneFlushInterval = pruneFlushInterval;
    -    this.pruneUpperBound = new AtomicLong();
    -    this.shouldFlush = new AtomicBoolean(false);
    -    startFlushThread();
    +    this.pruneEntries = new ConcurrentSkipListMap<>();
    +  }
    +
    +  public void persistPruneEntry(byte[] regionName, long pruneUpperBound) {
    +    // The number of entries in this map is bound by the number of regions in this region server and thus it will not
    +    // grow indefinitely
    +    pruneEntries.put(ByteBuffer.wrap(regionName), pruneUpperBound);
       }
     
       public boolean isAlive() {
    -    return flushThread.isAlive();
    +    return flushThread != null && flushThread.isAlive();
       }
     
    -  public void persistPruneEntry(long pruneUpperBound) {
    -    this.pruneUpperBound.set(pruneUpperBound);
    -    this.shouldFlush.set(true);
    +  @Override
    +  protected void startUp() throws Exception {
    +    LOG.info("Starting PruneUpperBoundWriter Thread.");
    +    startFlushThread();
       }
     
    -  public void stop() {
    +  @Override
    +  protected void shutDown() throws Exception {
         if (flushThread != null) {
    +      stopped = true;
    +      LOG.info("Stopping PruneUpperBoundWriter Thread.");
    --- End diff --
    
    It would be good to move this log message outside the if condition


> Share PruneUpperBoundWriter across all TransactionProcessors on the same region server
> --------------------------------------------------------------------------------------
>
>                 Key: TEPHRA-215
>                 URL: https://issues.apache.org/jira/browse/TEPHRA-215
>             Project: Tephra
>          Issue Type: Improvement
>    Affects Versions: 0.11.0-incubating
>            Reporter: Gokul Gunasekaran
>            Assignee: Gokul Gunasekaran
>             Fix For: 0.11.0-incubating
>
>
> Currently we start one prune upperbound writer thread per TransactionProcessor coprocessor. Instead we should be able to share only one thread that flushes writes to the prune state table periodically.


--
This message was sent by Atlassian JIRA
(v6.3.15#6346)